Please, use this thread to continue the extremely productive discussion we started here.
GDS voting
Thu, 04/30/2009 - 15:16
Thu, 04/30/2009 - 15:26
#1
My thoughts on the voting
Well, now that the results are in, it's a lot easier to see how the voting worked, and for the most part, I think that it gave helpful results.
I do think that it would be nice, in the future, to be able to abstain from choosing one or two of the places, ie; vote only for 1st and 2nd place if you feel that no games meet the criteria well enough to chose as a 3rd.
Many thanks to SEO for taking the time to get this discussion going and organize this month's GDS challenge!
Thu, 04/30/2009 - 16:16
#2
The voting seemed to work
The voting seemed to work fine. Only question I'd have is what's special about 5/4/3 and not something like 3/2/1 (since that's the kind of question i'd ask about scoring in a game design...)?
Thu, 04/30/2009 - 17:33
#4
Ranked Pairs
The Golden Geek Awards over on BGG each year uses the ranked pair method. (Wikipedia article: http://en.wikipedia.org/wiki/Ranked_pairs) Voters simply list the games they want to vote for in order, most favored first. Voters can simply omit any candidates they don't want to vote for, and ranked pairs also allows you to give equal ranks to entries. The result is a ranked list with the winner at the top. As a pure voting method, I'm a big fan of ranked pairs where you want a ranked result as the output.
A disadvantage is that there isn't a "vote tally" we can look at -- I'm not sure if that matters to everyone? In other words, if there were 8 entrants to the GDS, people could rank/vote for as many or as few as they wanted, and the result would be those 8 entries sorted from most popular to least popular, but we would not see a specific number of votes for each (well, we could list that too, but that's not as meaningful in the final result since it depends how each voter ranked the entries).
Sat, 05/02/2009 - 10:01
#5
Ranked Pairs
Excellent suggestion! At least at first sight, it takes care of many of the things mentioned in the previous discussion:
"two votes for second of third should beat just one for first"
"I want to rank all the games"
"I DON'T want to rank all the games"
etc.
I'm willing to give it a try in the near future. Looks like a bit more work for the admin than other methods, but I'm sure there has to be an application somewhere to automate the calculations. I failed to find one so far, but only did a quick search. If anyone has a link, please post it. If we can't find one done, I'm sure we can create one, if we decide to go with Ranked Pairs as the preferred GDS voting method.
I would love to know what the rest of you think of this.
Sat, 05/02/2009 - 10:54
#6
Actually...
You could provide a tally, and the tally would actually be rather large:
"Tally
To tally the votes, consider each voters' preferences. For example, if a voter states "A > B > C" (A is better than B, and B is better than C), the tally should add one for A in A vs. B, one for A in A vs. C, and one for B in B vs. C. Voters may also express indifference (e.g., A = B), and unstated candidates are assumed to be equally worse than the stated candidates.
Once tallied the majorities can be determined. If "Vxy" is the number of Votes that rank x over y, then "x" wins if Vxy > Vyx, and "y" wins if Vyx > Vxy."
So you could still present a tally, it just wouldn't be vote for vote.
I like it, we could try it next time. Maybe there is an excel function for it?
Sat, 05/02/2009 - 11:03
#7
Please keep in mind that any
Please keep in mind that any voting system discussed and *liked* should be programmable since I am in the middle of developing an automated GDS system for the site.
So as long as that voting system can be turned into an easy to display, easy to select/vote and easy to calculate, I would greatly appreciated it!
The GDS module I am creating will hopefully help to reduce the admin and user submissions process, in hopes to make GDS a manageable part of BGDF
Sat, 05/02/2009 - 11:35
#8
I found a calculator for
I found a calculator for Ranked Pair voting (among other methods) here.
Checked the results based on the voting for this month, and the results (for a rank produced by running successive calculations removing winners from the ballots) were:
Escape from Hadron
Hungry Cities
Blobs!!!
TIE: cAMP Purpureum, Ameban Amok
Altania
Ballo dei Pagliacci (Dance of the Clowns)
TIE: Tower, Firewall
The Duchies
So, basically pretty close to the current results (which I guess is a good sign). Obviously, had the Ranked Pairs method been decided, the votes could have been different.
Note to Zzzzz: Ranked Pairs is programmable (I used a programmed calculator), and it shouldn't be too hard to display the results as a ranking. Voting could be done either by entering a ballot in the "standard" format: A>D>F=I>G through an input box in a form, or by selecting from select lists. The drawback with select lists would be that while they work fine for >, we would need to provide a way to express ties. So I think the standard format in an input box would be just fine. Another option would be a list of all the candidates with a value field or select list with values for each.
Sat, 05/02/2009 - 14:57
#9
I liked the voting 1,2,3.
I
I liked the voting 1,2,3.
I think ranked pairs would work if it were programmed, with a fairly visual interface. kind of like how you can click arrows in your BGG top ten to move them up and down. Ranked pairs makes it a little more work on the voters, but I think if you are voting on GDS you probably can hack it.
Sun, 05/03/2009 - 03:25
#10
Maybe Aldie would share the
Maybe Aldie would share the code they use for the Golden Geek voting over on BGG. IIRC, the form displays all the candidates, initially with no rank next to them, and then you just type in a number next to each that you wish to vote; enter the same number for those you prefer equally.
Is it me or is the tally the same as ranked pairs?
Sun, 05/03/2009 - 09:18
#11
SiddGames wrote:you just type
you just type in a number next to each that you wish to vote; enter the same number for those you prefer equally.
Is it me or is the tally the same as ranked pairs?
If I'm reading your question properly, it is not, because if you vote A=10, B=1 and I vote B=10, A=9, a simple addition would result in A=19 (winner), B=11, while with ranked pairs the values are used only to set the relative ranking, so the result would be a tie (A>B, B>A).
What's good about ranked pairs is that it removes the fan-factor: I can't push my favorite by giving it 10 points and none to the rest, it's the same as giving it 1 point, it only sets it a step above the rest. Which, in some situations, might be considered a bad thing, instead of a strength (since I can't make my vote reflect that I may think that one game is far superior than the rest). But I guess you can't have both sides. :)
Mon, 05/04/2009 - 11:27
#12
I thought that's what Dylan's
I thought that's what Dylan's description of tally was - just showing preference, and not weighted votes...
Mon, 05/04/2009 - 11:43
#13
You're right
Yeah, sorry. I thought you were talking about a tally of the points awarded to each game. My bad.
Wed, 01/20/2010 - 18:20
#14
I'm back in this dead horse
I'm back in this dead horse of a thread to discuss Sedj's 6 votes vs 10 votes.
This month's GDS had a 6 vote, 3 max on any one entry instead of the 10 votes 5? max that we usually use.
So this month my voting looked like this:
Game A: 3 votes
Game B:1 vote
Game C:1 vote
with one vote unused, given the weight of the votes I couldn't find a 4th place game that I could give a vote to without hurting B and C
This is how my votes would have looked with 10 votes:
Game A: 3
Game B: 2
Game C: 2
Game D: 1
Game E: 1
Game F: 1
So the fewer votes had a pretty drastic effect on my voting, the game I liked best is now 3x as good as a second best game, as opposed to 3:2. And the third tier games where I liked something about them or they had a neat spark in them, get no votes vs 1s.
I don't think this is bad or good, just interesting.
Fri, 01/22/2010 - 10:09
#15
Complete listing
Would others like to find out how many votes their entries got?
If not, can I get a PM that tells me how I placed?
Fri, 01/22/2010 - 11:44
#16
Most people detail their
Most people detail their votes when they post their critiques. And Seth can probably post a detail of the votes each game got (keeping the identity of the voters secret, in case someone doesn't want to make his particular votes public).
Fri, 01/22/2010 - 12:46
#17
Votes are posted in the subject
If you note, the subject line of each entry now shows the game author and the number of votes it received.
Fri, 01/22/2010 - 13:18
#18
I saw that, I was thinking of
I saw that, I was thinking of adding detailed info like:
7 votes (3,2,1,1)
Not that the votes are actually of paramount importance; I rather have one detailed critique like stubert's than lots of detailed numeric information on the votes.
It's just a matter of relative value for each place.
With 5/4/3, two votes for third beat only one vote for first (so we're rewarding average popularity over small group of hardcore fans).
I don't mean that's necessarily better than the opposite, just that as the discussion progressed, most people seemed to agree that two votes for third should have a small advantage over just one vote for first.
Just checked the votes for this month, and -somehow ironically, given all this voting-method discussion going on-, 3/2/1 would have been the same as 5/4/3 was, or 5/3/1 would have been too. But we're game designer and want to refine our scoring systems, and take into account those not-so-common cases where one entry might get a lot of 2nd or 3rd place votes, but not so many 1st place ones. :)